2:30 PM - 2:45 PM
[SCG52-04] Building high-resolution oceanic crustal structures using full waveform inversion of long-offset active-source ocean bottom seismic data
Keywords:full waveform inversion, oceanic crust, ocean bottom seismometer, turning wave, seismic imaging
Travel-time tomography has been widely used for obtaining oceanic crustal models from marine seismic data recorded by ocean bottom seismometers (OBS) or multi-channel streamers (MCS). However, the high-frequency ray approximation limits its resolution, therefore it has difficulties in delineating fine-scale structures of the heterogeneous crust and upper mantle, for example, in the slow-spreading environments and in the subduction zones. On the other hand, full waveform inversion (FWI), which better captures the physics of seismic wave propagation, has received wide interest in recent years across industry and academia, thanks to the increasing capability of high-performance computing and the improvement of seismic data quality.
Here we applied the cutting-edge FWI technique to the crustal turning waves (Pg) recorded by the OBS spaced at 10-20 km from the slow-spreading equatorial Atlantic Ocean. The workflow includes a sequential application of the trace-normalized and the classic true-amplitude FWI, that match the phase and complete waveform information, respectively, allowing a smoothing transition from the tomography-derived starting model to model update including full physics. At each stage, the near-to-intermediate offset data is matched first before moving to the long-offset part, effectively updating the P-wave crustal model from shallower to deeper depth, because the turning waves of smaller offsets propagate in the shallower oceanic crust. In the survey perpendicular to the Mid-Atlantic Ridge covering 7-12 Ma old oceanic crust, the velocity model from FWI contains alternate 400-500 m thick layers with ±100-200 m/s P-wave velocity variations extended over 5-15 km in the lower crust. The layered anomalies in the lower crust provide evidence that the lower crust was formed in situ by melt sill injection, cooling, and crystallization at different depths. On a second OBS profile between Romanche and St. Faul transform fault zones in the Atlantic, the crustal velocity model exhibits strong heterogeneity that is blurred in the tomographic model, with four distinct 20-30 km long crustal segments which have a good correlation with the overlying seafloor morphology features. These results demonstrate the capability of the advanced FWI technique for determining high-resolution quantitative velocity models of the oceanic crust and could be applied to other similar data from different oceanic settings, to transform our understanding and to gain new knowledge on the formation and evolution of oceanic lithosphere.
Here we applied the cutting-edge FWI technique to the crustal turning waves (Pg) recorded by the OBS spaced at 10-20 km from the slow-spreading equatorial Atlantic Ocean. The workflow includes a sequential application of the trace-normalized and the classic true-amplitude FWI, that match the phase and complete waveform information, respectively, allowing a smoothing transition from the tomography-derived starting model to model update including full physics. At each stage, the near-to-intermediate offset data is matched first before moving to the long-offset part, effectively updating the P-wave crustal model from shallower to deeper depth, because the turning waves of smaller offsets propagate in the shallower oceanic crust. In the survey perpendicular to the Mid-Atlantic Ridge covering 7-12 Ma old oceanic crust, the velocity model from FWI contains alternate 400-500 m thick layers with ±100-200 m/s P-wave velocity variations extended over 5-15 km in the lower crust. The layered anomalies in the lower crust provide evidence that the lower crust was formed in situ by melt sill injection, cooling, and crystallization at different depths. On a second OBS profile between Romanche and St. Faul transform fault zones in the Atlantic, the crustal velocity model exhibits strong heterogeneity that is blurred in the tomographic model, with four distinct 20-30 km long crustal segments which have a good correlation with the overlying seafloor morphology features. These results demonstrate the capability of the advanced FWI technique for determining high-resolution quantitative velocity models of the oceanic crust and could be applied to other similar data from different oceanic settings, to transform our understanding and to gain new knowledge on the formation and evolution of oceanic lithosphere.